Accounts & Billing

Accounts, Billing & FAQ

URL: /docs/accounts-billing/

Manage your InstantGPU account, API keys, deployments, and account settings from your dashboard.

Account Management #

From your central InstantGPU dashboard, you can configure personal and organization settings, manage user access, configure SSH public keys, monitor active workloads, and review real-time resource utilization.

Profile & Security

Configure password-less login, Two-Factor Authentication (2FA), and authorized email addresses.

Teams & Organizations

Invite team members with role-based access control (Admin, Developer, Billing viewer).

SSH Key Management

API Keys #

API keys allow applications, CI/CD pipelines, and microservices to authenticate programmatically with InstantGPU APIs.

Security Notice

Keep API keys private and never expose them in publicly accessible client-side code, GitHub repositories, or browser bundles. Always store them as server-side environment variables (e.g. INSTANTGPU_API_KEY).

Billing #

InstantGPU billing depends on the service and compute configuration you use. Before launching a workload, review the applicable pricing and resource configuration.

GPU Cloud Billing

GPU Cloud workloads are billed according to the applicable GPU and compute configuration on an exact per-minute or per-hour basis while the instance is running.

Serverless Billing

Serverless workloads use usage-based GPU execution billed per second according to the applicable pricing model. Scale-to-zero workloads incur zero compute cost when idle.

Endpoint Billing

AI Endpoint usage is charged according to the applicable endpoint pricing and usage model (e.g. per generated image, per audio minute, or per 1K output tokens).

Frequently Asked Questions #

Find immediate answers to frequently asked questions regarding InstantGPU compute, models, and setup:

What is InstantGPU?

InstantGPU is a GPU infrastructure platform for AI workloads, providing GPU Cloud, Serverless GPU, AI Endpoints, and related infrastructure tailored for high-performance machine learning.

Which GPU should I choose?

Choose based on your model's VRAM requirements, performance requirements, workload type, and budget. For instance, RTX 4090 (24GB) is optimal for image diffusion and 8B LLMs, whereas A100 (80GB) and H100 are recommended for 70B+ LLMs and heavy model fine-tuning.

Can I deploy my own model?

Yes, where supported by the selected deployment method, you can use your own model and environment. You can pull weights from Hugging Face Hub, load weights from persistent storage volumes, or package them into containers.

Can I use Docker?

Yes, custom container workflows can be used where supported. Bring your own Docker container from Docker Hub, GitHub Container Registry, or private registries to run custom environments.

Can I access my GPU through SSH?

GPU Cloud environments can provide direct machine access through supported connection methods. Simply add your public SSH key to your account and connect directly via your terminal or IDE.

Can I use InstantGPU for production?

Yes. Choose the infrastructure and configuration appropriate for your production workload, traffic, scaling, and reliability requirements. InstantGPU supports health checks, autoscaling, and SLA-backed multi-region availability.

Get Started with InstantGPU → Create Account →